Food Insecurity and Poor Health – Unravelling the Mechanisms, Informing Better Solutions

Food insecurity—defined by the USDA as a household-level economic and social condition of limited or uncertain access to adequate food¹—has been a key public health metric in the U.S. and globally for several decades. Food insecurity is consistently associated with worse physical and mental health outcomes and health disparities, in particular due to cardiometabolic diseases, and is a foundational social determinant of health.² At the same time, many assumptions have been made about the mechanisms underlying these associations, and consequently, the corresponding best interventions to reduce the health and equity harms associated with food insecurity. For example, it’s commonly presumed that adverse impacts of food insecurity are predominantly mediated through worse nutrition due to insufficient purchasing power to obtain healthful foods; and therefore that increasing purchasing power for food will improve nutrition and thus health. However, empiric evidence demonstrates these assumptions to be overly simplistic.

Because of the tools and metrics used to assess it, food insecurity is predominantly a measure or proxy of poverty and financial strain. Accordingly, the mechanisms linking food insecurity to poor health and disparities are closely linked to the many challenges of low socioeconomic status, such as job opportunities, housing, neighborhood environment, education, transportation, community cohesion, substance abuse, and healthcare access.² Understanding how different tools capture food insecurity and its health consequences, as well as the potential mediating mechanisms, is relevant to designing effective clinical, public health, and policy interventions to address it.

In this issue, Tian and colleagues investigate how standard 10-item and shorter 6- and 2-item food security instruments predict premature all-cause and cardiovascular mortality in the National Health Interview Survey (NHIS), with replication in the National Health and Nutrition Examination Survey (NHANES). Importantly and as the most novel contribution, they also evaluated potential mediating factors in NHANES, providing inference on plausible pathways of effects.³

As expected, in crude (bivariate) analyses, many differences were identified among individuals living in food secure vs. insecure households. The latter were more likely to be younger, identify as non-Hispanic Black or Hispanic, have lower education and income, be unemployed, lack regular healthcare access or private health insurance, not own a home, or be unmarried or living without a partner. Individuals experiencing food insecurity were also more likely to be current smokers, not sleep 6–8 hours nightly, have lower physical activity, and have worse quality diets. About 2 in 5 reported receiving SNAP benefits in the last year.

In both datasets, after adjusting for sociodemographics, other social determinants of health, and smoking, the researchers found that household food insecurity defined by all three instruments significantly and similarly predicted all-cause mortality (~18–26% higher risk) and cardiovascular mortality (~27–50% higher risk). Interestingly, further adjustments for SNAP enrollment in NHIS; and for SNAP enrollment, diet quality, and total energy intake in NHANES; had little effect on these risk estimates. This suggests these factors, at least jointly considered, are not major mediators of the relationships between food security and all-cause or cardiovascular mortality. In comparison, further adjustment for cardiometabolic risk markers (i.e., adiposity, hemoglobin A1c, blood pressure, blood cholesterol), medication use, sleep, and depression attenuated the risk estimates by ~40%, suggesting this combination of factors partly mediate the observed higher risk associated with food insecurity.

Evaluation of mediating factors in groups can obscure various effects or even opposing effects of individual factors. When potential mediating factors were individually considered in quantitative mediation analysis, the largest identified relative contributors for the observed higher all-cause mortality were higher prevalence of diabetes (estimated to contribute to 10.9% of the risk difference), receiving SNAP (9.3%), low diet quality (8.9%), and suboptimal sleep duration (8.3%), and depression (4.1%). For cardiovascular mortality, the largest positive mediators (i.e., potentially explaining the harm) were low diet quality (10.9%), diabetes (9.6%), and depression (8.6%), obesity (6.5%), and suboptimal sleep duration (4.5%); while a large negative mediator (i.e., potentially counteracting the harm) was receiving SNAP (−19.3%). As discussed by the authors, in a few prior studies that considered potential mediators of associations between food insecurity and poor health, obesity, poor lifestyle, and poor mental health were identified as the most important mediators.

This emerging body of evidence has important implications for strategies to address food insecurity and its associated adverse health outcomes. First, given the significant associated health risk, both healthcare- and population-based efforts should regularly screen for food insecurity as a foundational predictor of well-being. Second, for both all-cause and cardiovascular mortality, the key identified mediators included diabetes, diet quality, mental health, and sleep. Thus, healthcare, community, and policy strategies to help improve the lives of those with food insecurity must consider not only food support programs, but also integrated approaches for diabetes prevention and management, improved nutrition security and diet quality, screening and support for mental health, and sleep evaluation and counseling.

The efficacy of SNAP, by far the largest federal nutrition program, in supporting improved health outcomes is controversial. Participation in SNAP reduces food insecurity and financial strain, and thereby also appears to have some benefits for mental health. At the same time, participation in SNAP, compared to income-eligible non-participants, is associated with worse diet quality and higher rates of cardiometabolic disease and mortality.^4–6 The present investigation by Tian and colleagues provides further conflicting evidence on the potential effects of SNAP. For cardiovascular mortality, SNAP participation appeared to significantly decrease the adverse impacts of food insecurity, suggesting a protective effect. Yet for all-cause mortality, SNAP participation appeared to significantly mediate the adverse impacts of food insecurity, suggesting harm. Research suggests that increasing the nutritional focus of SNAP through incentives for more healthful foods, taxes or restrictions on sugar-sweetened beverages and other unhealthful items, or a combined program of incentives plus disincentives (without restrictions) would each significantly improve the health of SNAP recipients and contribute to healthcare cost savings.^7,8

The U.S. and globe face a nutrition crisis, with diet-related diseases being a top threat to health, health equity, and economic solvency. In very recent years, the recognition of these threats, and the interest in addressing them, have dramatically increased among the public, healthcare, and state and federal government including both Congress and the White House. We are at a special moment in time to capture and leverage this recognition and interest to finally begin to address this crisis of poor nutrition and diet-related disease, in particular among those facing potent social determinants of health including food insecurity.^{2, 9}